Add periodic checking of stream/consumers
wallyqs opened this issue · comments
When a stream/consumer ends up in Error state due to an API call failing, the actual JS might actually be fine so need to retry when that happens to clear the Error state from the CRD.
if i understand this right, then this periodic checking would also solve the issue where a newly added (qa/review) controller currently does not pick up existing streams or consumers (it only looks for newly created ones)
Could this also fix the problem we have? we are currently in a testing phase and sometimes our VMs with k3s gets restarted, after the restart all services etc. are running again but we are missing our jetstreams that we created and we have to create them manually again or redo the whole deployment...
so we have the creation of the jetstream via a helm templates file and it obviously gets executed while deployment but not after a restart. Or do we explicitly have to create the streams via kubectl apply?
I'm wondering if you would be open to a PR refactoring the code to use https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.14.5 ?
This would also assist in running retries/continuous reconciliation. https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.14.5/pkg/reconcile#Result